A method for generating a stereophonic residual in the prior art includes: downmixing a signal input by a first sound channel with a signal input by a second sound channel to obtain a monophonic signal and a sideband signal; encoding the monophonic signal by using a monophonic encoding method; decoding the encoded signal to obtain a local decoded signal of the monophonic signal; extracting a stereophonic parameter from the signal input by the first sound channel and the signal input by the second sound channel, where the stereophonic parameter reflects a ratio of energy of the first sound channel to energy of the second sound channel; generating a sideband predication signal by using the local decoded signal and the stereophonic parameter; generating a sideband residual signal according to the sideband signal and the sideband predication signal; and then encoding the monophonic signal and the sideband residual signal.
In the method for generating a sideband residual signal in the prior art, the signal input by the first sound channel and the signal input by the second sound channel, which may be obtained by decoding at a decoding end, are related to the local decoded signal and the sideband residual signal. However, the local decoded signal and the sideband residual signal are encoded signals, and a quantization error exists in an encoding process. This quantization error is evenly allocated to the signal of the first sound channel and the signal of the second sound channel. When there is a larger difference between the energy of the signal of the first sound channel and the energy of the signal of the second sound channel of the stereophonic signal, a monophonic quantization error has a greater impact on a channel of signal whose energy is smaller, which reduces the quality of a signal that is generated according to the sideband residual signal.